brse-workflow-chainlisted

# BrSE Workflow Chain Use this skill as a meta-orchestrator. It does not replace the domain skills — it tells the BrSE which skill to invoke next and what context to carry between steps. ## When To Use - New customer request just arrived and BrSE does not yet know the shape of work. - A vague Japanese spec needs to be turned into dev-ready tickets and a customer reply in one session. - BrSE wants to ensure no step (clarify / impact / split / QA / report) is skipped. ## When NOT To Use - The needed work is a single skill in isolation — invoke that skill directly, do not over-orchestrate. - The BrSE is mid-chain in a previous run and wants to continue — resume the active chain instead of restarting. - The input is so small that running the chain costs more than the value (e.g., a one-line status reply) — invoke `brse-client-report` only. - The task is meta-work on the skill set itself — use `brse-skill-author`, not the chain. ## Classification Flowchart ```dot digraph workflow_chain { rankdir=TB; node [shape=box, style=rounded]; start [label="Customer input arrives", shape=oval]; q0 [label="Vague / hedged / short JP?", shape=diamond]; s0 [label="Stage 0\nbrse-intent-reader"]; q_type [label="Classify input type", shape=diamond]; feat [label="New feature request\n→ 1, 2, 3, 4, 5, 7"]; bug [label="Bug investigation\n→ 1, 3, 6, 7"]; spec [label="Spec change\n→ 1, 2, 3, 4, 7"]; status [label="Status reply\n→ 6, 7"]; branch_q [lab